1

I have a strange problem. After upgrading to jboss 6.1 final AS 7 the hornet queue listeners are firing after a delay of 1 day. I can see the queue has data which is more than the max limit so the data is being paged by jboss. But why the delay in consuming the messages?

Any help guys...

Adhir
  • 203
  • 1
  • 2
  • 13
  • I don't have any relevant information to help you... it could be an issue in your application still... Take a look on the print-data and print-pages if that brings you any information about the data you have. – Clebert Suconic Jun 24 '13 at 15:32
  • Could be an application specific problem, but why was it working before and stopped working after the upgrade. @Clebert, I know that you are actively involved with JBOSS support, would you be by any chance aware of any such issues, or are there similar issues reported fro JBOSS AEP 6.1 Final – Adhir Jun 26 '13 at 10:12
  • I am not aware of any similar issues. I would need to evaluate configuration, journal sizes, min pages.. max pages... There is actually just one thing, maxSize should be greater than pageSize. In an initial version this was the same in a few initial versions. If you provide me some data through JBoss Forums or support portal I can provide you some information, but I don't have much to say based on just this. – Clebert Suconic Jun 27 '13 at 02:51
  • we have the max size configured as 10 times the page size. And the problem is that this issue occurs intermittently in production environment. We have not been able to replicate it in a dev setup. Even if we hold the mdb using a breakpoint and fill the queue to its maximum capacity. Let me kno if this is enough info or I need to post something else. – Adhir Jul 09 '13 at 04:36
  • Do you have any redelivery delay set? Maybe there's a misconfiguration there – Clebert Suconic Jul 09 '13 at 15:24
  • You have EAP, right? you have a support contract. I would start a support ticket where you could provide a bit more data to support engineers or myself. – Clebert Suconic Jul 09 '13 at 15:25
  • That should do Clebert. Thank you. I will check the re-delivery delay and post it here or the ticket that you would create. We did hear from the JBOSS support, here is the extract -- "The only situations I can recall where messages are held for a prolonged period involve failover to backup broker that does not have consumers attached, if this happens messages can remain with the backup broker until the original broker is re-started and failback occurs, if you suspect that this may be happening you should be able to check the consumer count on the live destination using the CLI. " – Adhir Jul 11 '13 at 06:55

0 Answers0